U.S. patent number 5,322,573 [Application Number 07/955,512] was granted by the patent office on 1994-06-21 for inp solar cell with window layer.
This patent grant is currently assigned to The United States of America as represented by the Administrator of the. Invention is credited to Raj K. Jain, Geoffrey A. Landis.
United States Patent |
5,322,573 |
Jain , et al. |
June 21, 1994 |
InP solar cell with window layer
Abstract
The invention features a thin light transmissive layer of the
ternary semiconductor indium aluminum arsenide (InAlAs) as a front
surface passivation or "window" layer for p-on-n InP solar cells.
The window layers of the invention effectively reduce front surface
recombination of the object semiconductors thereby increasing the
efficiency of the cells.
Inventors: |
Jain; Raj K. (North Olmsted,
OH), Landis; Geoffrey A. (Brookpark, OH) |
Assignee: |
The United States of America as
represented by the Administrator of the (Washington,
DC)
|
Family
ID: |
25496918 |
Appl.
No.: |
07/955,512 |
Filed: |
October 2, 1992 |
Current U.S.
Class: |
136/252; 136/256;
257/434; 257/631; 257/76; 257/E31.022; 438/94 |
Current CPC
Class: |
H01L
31/03046 (20130101); H01L 31/0693 (20130101); Y02E
10/544 (20130101) |
Current International
Class: |
H01L
31/0304 (20060101); H01L 31/068 (20060101); H01L
31/06 (20060101); H01L 31/0264 (20060101); H01L
031/06 () |
Field of
Search: |
;136/252,256,249TJ,262
;437/5,938 ;257/76,434,631 |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
Aina et al, Appln. Phys. 53(17), pp. 1620-1622 (Oct. 1988). .
Landis et al., IEEE Transactions on Electron Devices, vol. 37, No.
2 (1990). .
R. K. Jain et al, Appl. Phys. Lett., vol. 59, pp. 2555-2557 (Nov.
1991). .
L. Aina et al, Appl. Phys. Lett., vol. 51, p. 1637 (Nov. 1987).
.
M. B. Spitzer et al, Conference Record, 21st IEEE Photovoltaic
Specialists Conf. (May 1990), pp. 196-206..
|
Primary Examiner: Weisstuch; Aaron
Attorney, Agent or Firm: Miller; Guy M. Mackin; James A.
Shook; Gene E.
Government Interests
The invention described herein was made in performance of work
under NASA Contract No. NAS-33-25266, and is subject to the
provision of 305 of the National Aeronautics and Space Act of 1958,
as amended, (42 U.S.C. .sctn. 2457).
Claims
What is claimed is:
1. An indium phosphide photo-receiving semiconductor device
comprising:
a) a first layer of p-type indium phosphide having a front surface
for receiving light;
b) a second layer of n-type indium phosphide forming a
semiconductor junction with said first layer;
c) a light transmissive window layer of p-type InAlAs disposed on
said front surface of said first layer of indium phosphide, said
window layer being from about 10 nm to about 100 nm thick, being
substantially lattice matched to said first layer of indium
phosphide and having a bandgap wider than that of said first layer
of indium phosphide, said window layer forming an energy barrier
that reduces the recombination of electrons and holes at said front
surface of said first layer of indium phosphide whereby the
efficiency of the device is increased.
2. The device according to claim 1, wherein the window layer is
In.sub.0.52 Al.sub.0.48 As.
3. The device according to claim 1, wherein said second layer of
InP is disposed on a supporting substrate comprising germanium or
silicon.
4. The InP device of claim 1 wherein the window layer is In.sub.1-x
Al.sub.x As where x has a value in the range from 0.48 to less than
1.00.
5. A method of improving the efficiency of a multilayer indium
phosphide semiconductor device having a first layer of p-InP and a
second layer of n-InP forming a semiconductor junction
therebetween, said first layer having a front surface for receiving
light, said method comprising reducing surface recombination by
forming a light transmissive window layer of InAlAs from about 10
nm to about 100 nm thick on said front surface of said first layer
indium phosphide, and which is substantially lattice matched to
said first layer of indium phosphide an having a wider bandgap than
said first layer of indium phosphide.
6. The method according to claim 5, further comprising doping said
the window layer as p-type.
7. The method according to claim 5, further comprising forming the
window layer of In.sub.0.52 Al.sub.0.48 As.
8. The method according to claim 5, further comprising forming said
second layer of InP on a supporting substrate comprising silicon or
germanium.
9. The method of claim 5 wherein said window layer is In.sub.1-x
Al.sub.x As where x has a value in the range from 0.48 to less than
1.00.
Description
BACKGROUND OF THE INVENTION
The invention relates to indium phosphide (InP) semiconductor
devices, and more particularly to a method of improving the
efficiency of InP solar cells and other photo-receivers, and the
improved cells produced thereby. InP solar cells have excellent
radiation tolerance and the potential for extremely high
efficiency. However, to date InP solar cells have not equalled the
efficiency of other materials such as GaAs because of its
propensity for high recombination of minority carriers at the front
surface of the cell. The purpose of this invention is to increase
the efficiency of InP solar cells by reducing the recombination of
minority carriers at the front surface.
In InP, one of the most significant causes of poor efficiency is
the recombination of electrons and holes at the front surface of
the device, a process called surface recombination. Experimentally
measured values of surface recombination velocity (SRV) on InP
solar cells are extremely high. The surface recombination velocity
calculated from measured values of the short wavelength quantum
efficiency and current-voltage characteristics of the best existing
InP solar cells is in the range of 5.times.10.sup.6 to 10.sup.7
cm/s, which is unfavorably high. Once recombined at the surface,
the carriers are lost and cannot contribute to either the output
voltage or the output current of the cell.
High front surface recombination in InP cells is typically
addressed by using n-on-p cell configurations, where n-type InP is
used as the front surface or "emitter" layer. High doping (n.sup.+)
levels in the n-InP and extremely shallow emitter layers contribute
to minimize surface losses in the emitter. However, the best n-on-p
type InP cells have 19% conversion efficiency measured at Air Mass
Zero (AM0), and the best p-on-n InP cells today have efficiencies
of 15% AM0. Thus, with the present technology, p-on-n cells are not
as efficient as n-on-p cells, and neither is as efficient as GaAs
cells, which have been manufactured with over 22% efficiency.
In front surface recombination velocity could be reduced,
theoretical studies by the inventors indicate that p-on-n cells
would have better open circuit voltage and higher efficiency than
n-on-p cells. The heavy doping of the emitter required in n-on-p
cells to minimize the adverse effects of high surface recombination
leads to efficiency loss. The shallow emitter required is difficult
to fabricate and has high resistance, leading to further
losses.
Some InP solar cells have been made using as the active material
InP which has been deposited or "grown" on a substrate of a
different crystalline material. This is advantageous because a
substrate may be used which is stronger, lighter or lower in cost
than single crystal wafers of InP, while retaining all of the
features of an InP solar cell. Two crystalline substrates of
particular interest are silicon and germanium, which are available
in wafer form. A difficulty in this approach, however, is that in
the process of growing the InP, the Si or Ge will tend to diffuse
from the substrate into the InP layer being grown, doping the InP
to n-type. The p-on-n type solar cell is preferable for this
application because it can make use of the n-type doping in the
base. Thus, reduction of surface recombination in p-on-n type InP
is particularly important for these cell designs.
InP solar cells can also be used as one element in a tandem solar
cell, where it is used in conjunction with other solar cell
materials. For example, an InP solar cell may be used on top of an
InGaAsP solar cell, so that the light which penetrates through the
InP can be usefully absorbed by the InGaAsP.
Finally, while the devices discussed in detail are solar cells,
other electronic devices designed to absorb light, such as
photodetectors, photodiodes, phototransistors, laser power
receivers, thermophotovoltaic cells and the like could also be
fabricated out of InP and would likewise benefit from reduced
surface recombination at the light absorbing surface. Hence, there
is a need for alternative methods of minimizing the effects of high
surface recombination velocities in InP that do not have the
disadvantages associated with current methods. By increasing the
efficiency of InP devices, and in particular p-on-n InP, in this
way, they will become more suitable for space power applications
where their excellent radiation tolerance can be exploited.
SUMMARY OF THE INVENTION
The invention provides an alternative means of overcoming the
adverse effects of high surface recombination velocities (SRV) in
InP without the disadvantages of prior methods. Moreover, the
invention makes possible the advantageous use of p-on-n InP in
solar cells and other light actuated devices or photo-receivers.
Surface recombination is reduced by putting an electrical energy
barrier between the place where the light is absorbed and the
surface of the InP, so that the carriers cannot get to the surface
to recombine. This can be accomplished by introducing a
heterojunction window layer that has a bandgap wider than the InP
on the front surface of the InP. Heterojunction layers to reduce
front surface recombination velocity have not been previously used
on high-efficiency InP solar cells.
The present invention features a thin light transmissive layer of
the ternary semiconductor indium aluminum arsenide (InAlAs) as a
front surface passivation or "window" layer for p-on-n InP
semiconductor devices, such as solar cells. The InAlAs employed in
the invention has a bandgap of about 1.465 eV, slightly wider than
that of InP, and is lattice matched to InP at a composition of
approximately In.sub.0.52 Al.sub.0.48 As. Experimental work
indicates that the InP/InAlAs heterojunction forms a "stagger"
alignment, with the discontinuity in conduction band energy greater
than the difference in band gaps. This means that InAlAs will form
a barrier to electrons, but a sink for holes. Hence InAlAs will be
effective for reducing SRV on p-type InP. The inventive InAlAs
window layer improves both the short circuit current density and
the open circuit voltage, yielding significant improvements in
efficiency as described in R. K. Jain and G. A. Landis, Appl. Phys.
Lett., 59(20), p. 2555 (1991), incorporated herein by
reference.
Wide bandgap heterojunction window layers, i.e., layers having a
bandgap significantly wider than the bandgap of the underlying
material, have been used on other semiconductor materials to reduce
surface recombination. For example, AlGaAs layers have been used on
GaAs solar cells. Since AlAs has an almost identical lattice
constant to GaAs, any composition of AlGaAs will be lattice matched
to a GaAs substrate and there is no need to use any particular
composition of AlGaAs. Moreover, the wide bandgap of AlGaAs, which
can be as much as 0.74 eV higher than that of GaAs for high
aluminum content material, means that AlGaAs can be used as a
window layer on either p-type or n-type GaAs.
By contrast, the bandgap of InAlAs is only marginally wider than
that of InP. The bandgap increase from InP to lattice-matched
InAlAs is only 0.115 eV, significantly less than the increase from
GaAs to AlGaAs. The low difference in bandgaps between InP and
InAlAs means that band-bending effects can be significant compared
to the change in bandgap energy. In fact, band-bending effects due
to the 0.294 eV discontinuity in the conduction band energies are
larger in energy than the barrier due to the increase in bandgap,
which means that a lattice matched InAlAs layer will not be
effective on n-type InP. The inventors have solved this problem by
using the InAlAs layer on p-type InP, where the discontinuity in
conduction band energy is favorable.
The low difference in bandgaps also means that InAlAs absorbs light
nearly as efficiently as InP, and hence thick InAlAs will not work
well as a window layer, which must be nearly transparent to light.
Hence, the effectiveness of InAlAs as a window layer is
significantly dependent on thickness.
InAlAs has been used in conjunction with InP for several other
types of semiconductor devices, including high-electron mobility
transistors (HEMTS), quantum wells, and heterojunction base
transistors. In these devices the InAlAs is typically used as a
layer in contact with InGaAs, and in most applications is not
doped. In these devices, the InAlAs is not used as a window layer
on the surface, is not designed to be light transmissive to
incident light, is usually either n-type or undoped rather than
p-type, and is not used to decrease surface recombination of
minority carriers.
While the invention is primarily aimed at improving InP solar
cells, it may be used in other applications as well, such as for
improving the efficiency of InP photodetectors and their
sensitivity to short wavelength light. The invention enables novel
high-efficiency p-on-n type InP devices to be made. This cannot be
achieved by methods known to the prior art. The solar cell of the
invention will be of higher efficiency than any other known
radiation-tolerant solar cell, and hence will have many
applications for use in space power.
In accordance with the foregoing it is an object of the invention
to provide an indium phosphide photo-receiving semiconductor device
comprising a first layer of indium phosphide having a front surface
for receiving light, a second layer of indium phosphide forming a
semiconductor junction with said first layer, and a light
transmissive window layer of semiconductor material disposed on
said front surface of said first layer of indium phosphide, said
material being substantially lattice matched to said InP and having
a wider bandgap than said InP. Preferably, the first layer of InP
is p-type and the second layer of InP is n-type and the window
layer is from about 10 nm to about 100 nm thick. Still more
preferably the window layer is p-type and selected from AlAsSb or
InAlAs.
A preferred embodiment of the invention is an indium phosphide
photo-receiving semiconductor device comprising a first layer of
p-type indium phosphide having a front surface for receiving light,
a second layer of n-type indium phosphide forming a semiconductor
junction with said first layer, and a light transmissive window
layer of InAlAs disposed on said front surface of said first layer
of indium phosphide. Preferably, the InAlAs is substantially
lattice matched to the InP and still more preferably is of the
composition In.sub.0.52 Al.sub.0.48 As and from about 10 nm to
about 100 nm thick. In yet another embodiment the second layer of
InP is disposed on a supporting substrate comprising germanium or
silicon.
In another embodiment there is provided a method of improving the
efficiency of a multilayer indium phosphide semiconductor device
having first and second layers of indium phosphide forming a
semiconductor junction therebetween, said first layer having a
front surface for receiving light. The method according to this
embodiment comprises reducing surface recombination by forming a
light transmissive window layer of substantially lattice matched
semiconductor material having a bandgap wider than said indium
phosphide on said front surface of said first layer of indium
phosphide. In a preferred method the window layer is formed of
InAlAs or AlAsSb to a thickness of from about 10 to about 100 nm
thick.
Still more preferably, the invention provides a method of improving
the efficiency of a multilayer indium phosphide semiconductor
device having a first layer of p-InP and a second layer of n-InP
forming a semiconductor junction therebetween, said first layer
having a front surface for receiving light, said method comprising
reducing surface recombination by forming a light transmissive
window layer of p-type InAlAs on said front surface of said first
layer of indium phosphide. Preferably the method comprises forming
the InAlAs substantially lattice matched to the InP and still more
preferably forming the window layer of In.sub.0.52 Al.sub.0.48 As
from about 10 nm to about 100 nm thick.
Many additional features, advantages and a fuller understanding of
the invention will be had from the following detailed description
of the preferred embodiments and accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a diagrammatic cross-sectional view of a device according
to the invention.
FIG. 2 is a graph of efficiency versus thickness of a cell
according to the invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
FIG. 1 shows the general configuration of a semiconductor device
constructed according to the invention. As shown, a solar cell 10
is comprised of a first layer 14 of p-type InP as the emitter and a
second layer 12 of n-type InP as the base. The first and second
layers of InP form a p/n semiconductor junction 13 at their
interface. The front surface 15 of the emitter 14 is the surface
upon which light is incident on the solar cell. The inventive
InAlAs window layer 16 is disposed on the front surface of the
emitter layer 14. The InAlAs window layer 16 is thin enough so as
to be substantially transparent to light in the solar spectrum.
Front surface current carrying contacts 20 are disposed on the
front surface 17 of the InAlAs window layer 16 and back surface
current carrying contacts 22 are disposed on the back surface 11 of
the base layer 12. The device shown includes a front surface
antireflection (AR) coating 18 disposed on the front surface of the
window layer 16.
It is to be understood that the form of the device depicted in FIG.
1 has been chosen only for the purpose of describing a particular
embodiment and function of the invention, and that variations in
the configuration of the semiconductor device will be apparent to
one of ordinary skill in the art in view of the instant disclosure.
For example, parameters such as the doping of the window, emitter,
and base layers may be varied to optimize the performance of the
cell and to maximize tolerance to radiation. The front surface
contacts may be made directly to the p-type InP rather than to the
InAlAs. The InAlAs layer may be "capped" by addition of a thin
layer of material on top of the InAlAs to prevent corrosion or
degradation of the InAlAs due to exposure to the environment. Of
course, if this is done the cap material must either be transparent
or else thin enough to avoid significant absorption of light.
Similarly, the thickness of the layers may be varied and the
invention may be used in conjunction with other improvements such
as a grooved or reflection-reduced front surface as is known in the
art.
The window layer 16 is a light transmissive semiconductor
substantially lattice matched to the InP. The semiconductor may be
selected from a family of quaternary semiconductors including
indium, aluminum, arsenic, and antimony, with the formula
In.sub.(1-x) Al.sub.x As.sub.(1-y) Sb.sub.y, where x and y are
greater than zero and less than one. In general, InAlAsSb
quaternary materials are more difficult to manufacture than ternary
materials and not all of them will have bandgaps wider than that of
InP. However, out of the InAlAsSb family, the ternary
semiconductors InAlAs and AlAsSb are preferred as window layers
because the value of the parameters x and y can be found such that
the compositions AlAs.sub.1-y) Sb.sub.y and In.sub.(1-x) Al.sub.x
As are lattice matched to InP and have bandgaps wider than that of
InP.
In the preferred embodiment the window layer 16 is formed of the
compound InAlAs, which is defined as any material of the family of
ternary semiconductors of the formula In.sub.(1-x) Al.sub.x As,
where x may be any amount greater than zero and less than one. As
the parameter x is varied, both the electronic bandgap and the
lattice parameter of the semiconductor will change. The window
layer 16 is lattice matched to the underlying InP emitter by
forming the ternary compound In.sub.(1-x) Al.sub.x As, where x is
approximately 0.48, whereby the crystal lattice of the InAlAs will
match that of the InP substrate. This is advantageous because if
the materials are lattice matched, i.e., have the same lattice
constant, then crystal growth of the InAlAs layer can take place
with a continuation of the same crystalline structure as the
underlying InP. This minimizes the number of defects that can form
at the InAlAs/InP interface and hence, the principal locations for
surface recombination. In the absence of lattice matching, defects
are generated at the interface due to the mismatch. These defects
allow for the recombination of electron-hole pairs, thereby
degrading the device performance.
Some amount of reduction of surface recombination is expected to
result with layers composed of any composition of InAlAs that has a
bandgap greater than the 1.35 eV bandgap of InP. The widest bandgap
in the InAlAs system is achieved with the highest Al content and is
about 2.16 eV, but this composition is not lattice matched to InP.
It is most desirable to form the InAlAs composition with a lattice
constant matched to that of InP. This composition has a bandgap of
approximately 1.46 to 1.47 eV. However, if the layer is
sufficiently thin, a composition with a lattice constant different
from the lattice constant of InP, i.e., a composition slightly
mismatched from InP, can be used without the deleterious
dislocations at the interface, by the process of strained-layer
epitaxy. The required thickness will depend upon how much the
lattice constant differs from that of InP. Thicknesses of about 10
nm are expected to be thin enough that strained-layer epitaxy will
be possible.
FIG. 2 shows the calculated improvement of conversion efficiency as
a function of InAlAs window layer thickness. The cell efficiency is
best for the thinnest InAlAs layers where the minimum amount of
light is absorbed in the window layer. Improvements of more than
30% are predicted for p-on-n InP cells having the inventive InAlAs
window layers thinner than 15 nm. The improvement disappears for
layers thicker than about 100 nm because losses due to light
absorbed by the InAlAs layer outweigh the increases due to the
lower effective SRV.
For the advantageous use of p-on-n InP the InAlAs window layer has
p-type doping. Typical p-type dopants are zinc and cadmium in
amounts ranging from about 10.sup.15 to about 10.sup.19 cm.sup.-3,
with the window layer preferably being doped to about 10.sup.18
cm.sup.-3.
The use of lattice matched p-type indium aluminum arsenide as a
window layer significantly improves the performance of p-on-n type
InP solar cells and other optoelectronic devices. The improvement
in the p-on-n cell is accounted for by the energy discontinuity at
the heterojunction. In the case of the InAlAs/InP heterojunction,
it has been measured that the discontinuity in the conduction band
energies, approximately 0.294 eV, is more than the difference in
the bandgaps of the two materials, e.g., about 0.115 eV. This
results in a staggered alignment of the band edges, which provides
an energy barrier to minority carrier electrons, and hence reduces
surface recombination by making it harder for the electrons to
reach the surface.
The InAlAs window layers lattice matched to InP may be grown by
methods known to those of ordinary skill in the art, such as
atmospheric pressure organometallic vapor phase epitaxy (OMVPE) as
described in L. Aina and M. Mattingly, Applied Physics Letters,
Vol. 51, No. 20, p. 1637 (1987). Using trimethylaluminum,
trimethylindium and arsine as sources of the aluminum, indium and
arsenic, respectively, the flow of growth gasses is carried in
hydrogen over the InP substrate heated by, for example, RF to the
desired growth temperature, e.g., about 650.degree. C. In order to
prevent loss of phosphorus from the InP substrate while it is being
heated before growth, a flow of phosphine over the substrate may be
desirable.
Typically, the ratio of the group V source to the group III source
flow rate during growth is about 60, although the optimum ratio
will vary depending upon growth conditions. The trimethylaluminum
and trimethylindium flow rates are adjusted to produce the desired
lattice matched growth, which can be measured using double-crystal
x-ray diffraction. P-type doping is achieved using a group II
source. A common p-type dopant is zinc, which is incorporated by
using diethylzinc as a source gas.
EXAMPLE 1
The modelled cell has a 0.15 .mu.m thick p-type InP emitter doped
to 10.sup.18 cm.sup.-3 and a 5.0 .mu.m thick n-type base doped to
10.sup.17 cm.sup.-3. The InAlAs window layer is 20 nm thick and has
a doping concentration of 10.sup.18 cm.sup.-3. The cell includes a
two-layer antireflection coating comprised of 50 nm of ZnS on 100
nm of MgF.sub.2 to reduce light reflection. The front surface
contact metallization yielded about 5% grid coverage.
For calculation of the cell parameters minority carrier diffusion
lengths of 0.5, 0.5, and 2.0 .mu.m were assumed for the window
layer, emitter, and base layers, respectively. The front and back
surface SRV was assumed to be 10.sup.7 cm/s. These values are
typical of the current state of the art in p.sup.+ n InP solar
cells. A value of 1.465 eV was used for the band gap of the lattice
matched In.sub.0.52 Al.sub.0.48 As based on reported values of from
1.46 to 1.47 eV. The discontinuity in the conduction band energy at
the heterojunction was 0.294 eV. The numerical code PC-1D was used
to solve the semiconductor device transportation equations. The
short circuit current density for this cell was 32.7 mA/cm.sup.2,
the open circuit voltage was 925.6 mV, and the conversion
efficiency was 18.74% (AM0).
For comparison, measurements were taken for an identical cell
without the InAlAs window layer. For this cell the short circuit
current density was 26.9 mA/cm.sup.2, the open circuit voltage was
892.7 mV and the efficiency was 14.7% (AM0). It is clear that the
cell efficiency improved significantly, with both the short circuit
current density and the open circuit voltage improving with the use
of the InAlAs layer.
The efficiency gain due to the window layer is even more
significant for higher performance cells, since the effect of
surface recombination is even more important to the performance of
these cells because in such cells other sources of recombination
are decreased. For a higher efficiency InP cell with diffusion
lengths of 2 .mu.m for the emitter and 5 .mu.m for the base without
a window layer, the conversion efficiency is only about 15.5%
(AM0), a modest increase over the average 14.7% of the baseline InP
cells. Adding a 10 nm InAlAs window layer to such a cell (assuming
a 2 .mu.m diffusion length) increased the efficiency to 23.0% AM0,
a considerable increase.
Many modifications and variations of the invention will be apparent
to those skilled in the art in light of the foregoing detailed
disclosure. Therefore, within the scope of the appended claims, the
invention can be practiced otherwise than as specifically shown and
described.
* * * * *